30 Jan AI-Supported Mammography Screening Showed Consistently More Favourable outcomes Compared with Standard Screening

MedicalResearch.com Interview with:

Dr. Lång

Kristina Lång MD PhD
Associate professor, Diagnostic Radiology
Translational Medicine, Lund University
Senior consultant, Unilabs Mammography Unit
Skåne University Hospital, Malmö, Sweden

MedicalResearch.com: What is the background for this study?

Response:  Prior to the start of the trial, several retrospective studies had shown that AI could discriminate between screening mammograms at low and high risk of cancer, with performance comparable to that of average breast radiologists. These findings suggested a potential to improve both the efficiency and sensitivity of mammography screening. This motivated us to design and evaluate an AI-supported screening procedure in a randomised controlled trial. The MASAI trial was among the first prospective studies in this field and, to date, remains the only randomised trial with reported results on the use of AI in breast cancer screening.

In European breast cancer screening programmes, every mammogram is usually read by two radiologists, so called double reading, to ensure a high sensitivity. In the MASAI trial we compared AI-supported mammography screening to standard double reading without AI. I

n the AI-supported approach, mammograms identified as low-risk by the AI were read by a single radiologist, while high-risk mammograms underwent double reading, with AI providing additional detection support.

MedicalResearch.com: What are the main findings?

Response:  AI-supported mammography screening showed consistently more favourable outcomes compared with standard screening, including a 29% higher cancer detection rate, without increasing false positives, and 12% fewer interval cancers.

Importantly, the findings suggest a shift toward earlier detection of clinically relevant cancers, reflected in fewer aggressive or advanced interval cancers. In addition, the use of AI reduced the screen-reading workload for radiologists by almost half compared with standard double reading. This reduction is particularly relevant in the context of a growing shortage of breast radiologists.

MedicalResearch.com: How difficult is it to implement AI-supported mammography within current radiology technology?

Response: Implementing AI in screening requires adapting the IT infrastructure so that mammograms can be analysed by AI before being reviewed by radiologists. The analysis is fast, can run at any time, and does not delay the screening process. After implementation, it is important to continuously monitor AI and screening performance.
Some regions in Denmark and Sweden are already successfully using AI in their breast cancer screening programmes, demonstrating its feasibility.

MedicalResearch.com: What should readers take away from your report?

Response: Taken together, these findings indicate that AI-supported mammography screening is more accurate and efficient than standard double reading.

Further studies in other screening programmes are needed to assess the generalisability of the MASAI findings, and several such studies are already underway.

MedicalResearch.com: What recommendations do you have for future research as a results of this study?

Response:  With respect to replication, our findings of higher cancer detection rate, low false-positive rate, and reduced workload are consistent with results from other prospective and observational studies. However, this is the first study to report interval cancer rates in the context of AI-supported mammography screening. Replication of this key measure of screening efficacy will therefore be particularly important to further strengthen the evidence base.

Additional randomised controlled trials are ongoing or about to be launched and are expected to contribute valuable data. For us, the highest priorities are extended follow-up of the MASAI trial and robust cost-effectiveness analyses, as these can be generated relatively efficiently and are most likely to provide evidence that is directly actionable for health-policy decision-making.

MedicalResearch.com: Is there anything else you would like to add? Any disclosures?

Response: It is important to determine whether the additional costs from the AI software is balanced by savings from earlier treatment of cancers and reduced screen-reading workload for radiologists. A cost-effectiveness analysis based on the trial is currently underway.

We have nothing to disclose.

Citation: Interval cancer, sensitivity, and specificity comparing AI-supported mammography screening with standard double reading without AI in the MASAI study: a randomised, controlled, non-inferiority, single-blinded, population-based, screening-accuracy trial

—

The information on MedicalResearch.com is provided for educational purposes only, and is in no way intended to diagnose, cure, or treat any medical or other condition.
Some links may be sponsored. Products are not warranted or endorsed. Always seek the advice of your physician or other qualified health and ask your doctor any questions you may have regarding a medical condition. In addition to all other limitations and disclaimers in this agreement, service provider and its third party providers disclaim any liability or loss in connection with the content provided on this website.

Last Updated on January 30, 2026 by Marie Benz MD FAAD